Dylan Patel — The single biggest bottleneck to scaling AI compute

Explains when roughly $600B of hyperscaler capex actually becomes usable compute and how it compares with lab raises. This informs expectations of future model capability, pricing and availability.

• videoRyan Greenblatt – What happens once AI can automate AI research?
• videoJeff Dean & Noam Shazeer — 25 years at Google: from PageRank to AGI
• videoDylan Patel – Two labs will soon control most of the world's workforce
• videoHow GPT, Claude, and Gemini are actually trained and served – Reiner Pope

• videoDylan Patel – Two labs will soon control most of the world's workforceDwarkesh Patel
• videoInside the Trillion-Dollar AI Buildout | Dylan Patel InterviewInvest Like the Best
• podcastDylan Patel — Deep dive on the 3 big bottlenecks to scaling AI computeDwarkesh Patel

Checking sign-in…

Loading comments…